Generating and testing hypotheses about the fossil record of insect herbivory with a theoretical ecospace

A typical fossil flora examined for insect herbivory contains a few hundred leaves and a dozen or two insect damage types. Paleontologists employ a wide variety of metrics to assess differences in herbivory among assemblages: damage type diversity, intensity (the proportion of leaves, or of leaf surface area, with insect damage), the evenness of diversity, and comparisons of the evenness and diversity of the flora to the evenness and diversity of damage types. Although the number of metrics calculated is quite large, given the amount of data that is usually available, the study of insect herbivory in the fossil record still lacks a quantitative framework that can be used to distinguish among different causes of increased insect herbivory and to generate null hypotheses of the magnitude of changes in insect herbivory over time. Moreover, estimates of damage type diversity, the most common metric, are generated with inconsistent sampling standardization routines. Here we demonstrate that coverage-based rarefaction yields valid, reliable estimates of damage type diversity that are robust to differences among floral assemblages in the number of leaves examined, average leaf surface area, and the inclusion of plant organs other than leaves such as seeds and axes. We outline the potential of a theoretical ecospace that combines various metrics to distinguish between potential causes of increased herbivory. We close with a discussion of the most appropriate uses of a theoretical ecospace for insect herbivory, with the overlapping damage type diversities of Paleozoic gymnosperms and Cenozoic angiosperms as a brief case study.